Genetic factors have been demonstrated in certain atypical forms of very early onset Alzheimer's disease (AD) and familial aggregation has been repeatedly found in clinically identified series of AD probands. Nevertheless, the presence of familial factors in very late onset (75+ years) AD, the most commonly manifested form of the illness, is largely unknown. Some evidence from previous genetic epidemiological investion and recent unpublished data from our center suggest that the degree of heritability of AD is associated wit the AD onset age in the proband. Yet several studies of AD cases have not, to date, consistently found that the risk of relatives for an AD-like primary progressive dementia (PPD) is associated with the age at onset in the AD proband. The inconsistencies in previous studies may ell be in part attributable to small proband samples and investigative strategies that have been relatively crude. Furthermore, the probands included in most of the previously published investigations have been ascertained through clinical settings. This ascertainment method will tend to yield samples of AD patients biased toward an early onset age and may include a disproportionate number of probands with a positive family history. The close association between Mt. Sinai School of Medicine and the Jewish Home and Hospital for the Age (JHHA) offers the opportunity to study the familial characteristics of both very late onset AD and earlier onset AD probands using family history assessment techniques. The recruitment core will refer all JHHA admissions and those in the psychogeriatric clinics of the Mt. Sinai, Elmhurst, Bronx VA yielding a well- characterized sample of probable AD probands, non-AD demented probands, and cognitively intact very elderly controls (from the JHHA population) for family history assessment. The families of unambiguously non- demented spouses of the referred patients, assessed through family informants, will also be evaluated. The cumulative risks and the age specific hazard rate of PPD-reflecting the patterns of incidence over the late lifespan- will be assessed in probable AD probands with different on set times and comparison groups, helping to clarify the role of onset in familial risk of PPD. The examination of patients with neuropathologically confirmed AD will eliminated one source of error, diagnostic misclassification of the proband, in the risk estimates. Possible biases for family history status between sites and between autopsied and non-autopsied probable AD probands will also be examined.